view release on metacpan or  search on metacpan

clib/libmpeg2/motion_comp_vis.c  view on Meta::CPAN

/*
 * motion_comp_vis.c
 * Copyright (C) 2003 David S. Miller <davem@redhat.com>
 *
 * This file is part of mpeg2dec, a free MPEG-2 video stream decoder.
 * See http://libmpeg2.sourceforge.net/ for updates.
 *
 * mpeg2dec is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * mpeg2dec is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
 */

#include "config.h"

#ifdef ARCH_SPARC

#include <inttypes.h>

#include "mpeg2.h"
#include "attributes.h"
#include "mpeg2_internal.h"
#include "vis.h"

/* The trick used in some of this file is the formula from the MMX
 * motion comp code, which is:
 *
 * (x+y+1)>>1 == (x|y)-((x^y)>>1)
 *
 * This allows us to average 8 bytes at a time in a 64-bit FPU reg.
 * We avoid overflows by masking before we do the shift, and we
 * implement the shift by multiplying by 1/2 using mul8x16.  So in
 * VIS this is (assume 'x' is in f0, 'y' is in f2, a repeating mask
 * of '0xfe' is in f4, a repeating mask of '0x7f' is in f6, and
 * the value 0x80808080 is in f8):
 *
 *	fxor		f0, f2, f10
 *	fand		f10, f4, f10
 *	fmul8x16	f8, f10, f10
 *	fand		f10, f6, f10
 *	for		f0, f2, f12
 *	fpsub16		f12, f10, f10
 */

#define DUP4(x) {x, x, x, x}
#define DUP8(x) {x, x, x, x, x, x, x, x}
static const int16_t constants1[] ATTR_ALIGN(8) = DUP4 (1);
static const int16_t constants2[] ATTR_ALIGN(8) = DUP4 (2);
static const int16_t constants3[] ATTR_ALIGN(8) = DUP4 (3);
static const int16_t constants6[] ATTR_ALIGN(8) = DUP4 (6);
static const int8_t constants_fe[] ATTR_ALIGN(8) = DUP8 (0xfe);
static const int8_t constants_7f[] ATTR_ALIGN(8) = DUP8 (0x7f);
static const int8_t constants128[] ATTR_ALIGN(8) = DUP8 (128);
static const int16_t constants256_512[] ATTR_ALIGN(8) =
	{256, 512, 256, 512};
static const int16_t constants256_1024[] ATTR_ALIGN(8) =
	{256, 1024, 256, 1024};

#define REF_0		0
#define REF_0_1		1
#define REF_2		2
#define REF_2_1		3
#define REF_4		4
#define REF_4_1		5
#define REF_6		6
#define REF_6_1		7
#define REF_S0		8
#define REF_S0_1	9
#define REF_S2		10
#define REF_S2_1	11
#define REF_S4		12
#define REF_S4_1	13
#define REF_S6		14
#define REF_S6_1	15
#define DST_0		16
#define DST_1		17
#define DST_2		18
#define DST_3		19
#define CONST_1		20
#define CONST_2		20
#define CONST_3		20
#define CONST_6		20
#define MASK_fe		20
#define CONST_128	22
#define CONST_256	22